Re: [OMPI devel] MPI_Bcast

2012-11-19 Thread Jeff Squyres
+1 on what George says. In general, the only MPI collective operation that *forces* an explicit, global synchronization is MPI_Barrier. Specifically: no process can exit the barrier until all processes have entered the barrier. In you example (MPI_Bcast), none of the non-root processes can exi

[OMPI devel] Dec. devel meeting

2012-11-19 Thread Jeff Squyres
We're finally starting to fill in agenda topics on the wiki page. Please also fill in your name on the wiki page if you plan to attend: https://svn.open-mpi.org/trac/ompi/wiki/Dec12Meeting I need to setup guest badges and wifi for everyone, etc., so I need your names ahead of time. Thanks!

Re: [OMPI devel] [EXTERNAL] Re: Compile-time MPI_Datatype checking

2012-11-19 Thread Jeff Squyres
(SC is now over, and, aside from the US holiday later this week, I'm picking up the items that I've let fall on the floor for the past 1-2 months...) On Oct 31, 2012, at 3:51 PM, Barrett, Brian W wrote: > That said, if we didn't throw a warning if the pointer is of type void* or > char*, I thin

Re: [OMPI devel] [EXTERNAL] Re: Compile-time MPI_Datatype checking

2012-11-19 Thread Dmitri Gribenko
On Mon, Nov 19, 2012 at 6:35 PM, Jeff Squyres wrote: > (SC is now over, and, aside from the US holiday later this week, I'm picking > up the items that I've let fall on the floor for the past 1-2 months...) > > On Oct 31, 2012, at 3:51 PM, Barrett, Brian W wrote: > >> That said, if we didn't thro

Re: [OMPI devel] [EXTERNAL] Re: Compile-time MPI_Datatype checking

2012-11-19 Thread Jeff Squyres
On Nov 19, 2012, at 11:47 AM, Dmitri Gribenko wrote: >> 3. the warnings are disabled for buffer types of (char*) or (void*) > > I think that disabling warnings for char* might make sense. But > during the LLVM developer meeting I talked with Hal Finkel (MPICH2 > developer) about this. These ann

Re: [OMPI devel] [EXTERNAL] Re: Compile-time MPI_Datatype checking

2012-11-19 Thread Dmitri Gribenko
On Mon, Nov 19, 2012 at 6:51 PM, Jeff Squyres wrote: > On Nov 19, 2012, at 11:47 AM, Dmitri Gribenko wrote: > >>> 3. the warnings are disabled for buffer types of (char*) or (void*) >> >> I think that disabling warnings for char* might make sense. But >> during the LLVM developer meeting I talked

[OMPI devel] RFC: make mca_base_param_deregister actually work

2012-11-19 Thread Nathan Hjelm
What: mca_base_param provides a deregister function. In the trunk this function is only used in a ft path: ompi-trunk hjelmn$ find . -name \*.[ch] | xargs grep mca_base_param_deregister ./ompi/mca/bml/r2/bml_r2_ft.c:mca_base_param_deregister(param_type); Right now this function is broken